Counting 2,129 Big Data & Machine Learning Frameworks, Toolsets, and Examples...
Suggestion? Feedback? Tweet @stkim1

Author
Last Commit
Oct. 11, 2017
Created
Jul. 4, 2017

AerialCrackDetection

AerialCrackDetection is a project about object detection from aerial imagery using pavament crack data. The project uses the open source software library Caffe, with a ZF or VGG neuronal network. AerialCrackDetection is based on Faster RCNN.

    PS: The project is only the original version, the improved version is not open.

First part : Collecting data

  • The first part is collecting and labeling aerial pictures.
  • Most of the pictures are from School of Aerospace Engineering, Beijing Institute of Technology.
  • You can use LabelImg to analyze them and label them.
  • You can find the AerialCrackDataset in my Google Drive.
  • If you find AerialCrackDataset useful in your research, please consider citing:
    @inproceedings{
        Author = {Bo Wang},
        Title = {AerialCrackDataset: Towards Object Detection with Dataset},
        Laboratory = {Key Laboratory of Optoelectronic Imaging Technology and System, 
                      Ministry of Education, School of Optoelectronics, 
                      Beijing Institute of Technology},
        Year = {2017}
    }

Second part : Installing and Configuration

  • You need install the Caffe.
  • You need downloaded the Pre-trained ImageNet models: ZF and VGG16.
  • Detailed installation process reference Faster R-CNN.

Third part : Training with Caffe

  • The third part is training the detection and classification model.
cd $FRCN_ROOT
./experiments/scripts/faster_rcnn_end2end.sh [GPU_ID] [NET] [--set ...]
# GPU_ID is the GPU you want to train on
# NET in {ZF, VGG_CNN_M_1024, VGG16} is the network arch to use
# --set ... allows you to specify fast_rcnn.config options, e.g.

Result

Detection-Example1

Detection-Example1